|本期目录/Table of Contents|

[1]周江兵,吕杭炳.基于树模型的虚拟测量算法在半导体制造中的应用[J].电子设计工程,2020,28(01):49-54.[doi:10.14022/j.issn1674-6236.2020.01.012]
 ZHOU Jiangbing,LV Hangbing.Application of virtual measurement algorithm based on tree model in semiconductor manufacturing[J].SAMSON,2020,28(01):49-54.[doi:10.14022/j.issn1674-6236.2020.01.012]
点击复制

基于树模型的虚拟测量算法在半导体制造中的应用(PDF)
分享到:

《电子设计工程》[ISSN:1674-6236/CN:61-1477/TN]

卷:
28
期数:
2020年01期
页码:
49-54
栏目:
计算机技术应用
出版日期:
2020-01-05

文章信息/Info

Title:
Application of virtual measurement algorithm based on tree model in semiconductor manufacturing
文章编号:
1674-6236(2020)01-0049-06
作者:
周江兵1吕杭炳2
(1.中国科学院大学 微电子学院, 北京 100000;2.中国科学院微电子研究所 北京 100000)
Author(s):
ZHOU Jiang?bing1LV Hang?bing2
(1. School of Microelectronics, University of Chinese Academy of Sciences, Beijing 100000, China;2. Institute of Microelectronics?, Chinese Academy of Sciences, Beijing 100000, China)
关键词:
虚拟测量 特征提取 随机森林 梯度提升
Keywords:
virtual metrology feature extraction random forest gradient boosting
分类号:
TP3
DOI:
10.14022/j.issn1674-6236.2020.01.012
文献标志码:
A
摘要:
半导体制造是一个大批量多阶段生产的系统,工艺技术复杂、工序步骤繁多,稍有不慎就可能使晶圆的表面和内部产生缺陷,从而影响生产的质量和效率。将半导体制造业和互联网技术相结合,可以提高生产制造的效率和质量,降低成本和损耗。本文将进行数据采集和预处理,从数据中选择并提取出重要特征,并使用树模型中随机森林和梯度提升树两种算法来实现W plug中对电阻率的虚拟测量。实验结果表明,本文提出的方法与传统算法相比,测量误差减少近50%,能够准确预测出生产过程中电阻率变化的趋势,捕获到异常点;在实际的生产环境之中可以及时的给生产线上报警。
Abstract:
Semiconductor manufacturing is a high-volume multi-stage production system with complex process technology and numerous process steps. A slight inadvertentity can cause defects on the surface and inside of the wafer, which affects the quality and efficiency of production. Combining semiconductor manufacturing with Internet technology can increase the efficiency and quality of manufacturing and reduce costs and losses. This paper will collect and preprocess the data, select and extract important features from the data, and use the random forest and gradient boosting tree algorithms in the tree model torealize the Virtual Metrology of resistivity in the W plug. The experimental results show that compared with the traditional algorithm, the proposed method reduces the measurement error by nearly 50%.It can also accurately predict the trend of resistivity change during production, and capture the abnormal point. In the actual production environment, it can be timely Alarm on the production line.

参考文献/References:

[1] 张渊.半导体制造工艺[M].北京:机械工业出版社,2015.[2] 韩郑生.芯片制造:半导体工艺制程实用教程[M].6版.北京:电子工业出版社,2015.[3] 宁永铎.基于大数据的硅片形状诊断与预报[D].北京:北京有色金属研究总院,2018.[4] 盖天洋,粟雅娟,陈颖韦,等.基于机器学习的光刻坏点检测研究进展[J].微纳电子技术,2019(6):421-428,434.[5] 段大高,盖新新,韩忠明,等.基于梯度提升决策树的微博虚假消息检测[J].计算机应用,2018,38(6):410-414,420.[6] 杨立洪,白肇强.基于二次组合的特征工程与XGBoost模型的用户行为预测[J].科学技术与工程,2018(14):186-189.[7] Chen, Tianqi,Guestrin, Carlos.XGBoost: A Scalable Tree Boosting System[C].The22nd ACM SIGKDD International Conference,2016.[8] Fan-Tien Cheng, Chi-An Kao, Chun-Fang Chen, et al.Tutorial on applying the VM technology for TFT-LCD manufacturing[J].IEEE Transactions on Semiconductor Manufacturing,2015,28(1):55 - 69.[9] Hendrik Purwins,Bernd Barak,Ahmed Nagi.Regression methods for virtual metrolo-gy of layer thickness in chemical vapor deposition[C]//IEEE/ASME Transactions on Mechatronics,2014:1-8.[10]Seokho Kang, Dongil Kim, Sungzoon Cho.Efficient feature selection-based on random forward search for virtual metrology modeling[J].IEEE Transactions on Semiconductor Manufacturing,2016,29(4):391-398.[11]Daniel Kurz, Cristina De Luca, Jürgen Pilz. A sampling decision system for virtual metrology in semiconductor manufacturing[J].IEEE Transactions on Automation Science and Engineering,2015,12(1):75-83.[12]Seokho Kang.On effectiveness of transfer learning approach for neural network-based virtual metrology modeling[J].IEEE Transactions on Semiconductor Manufacturing,2018,31(1):149-455.[13]James Moyne, Jamini Samantaray, Michael Armacost.Big data capabilities applied to semiconductor manufacturing advanced process control[J]. IEEE Transac-tions on Semiconductor Manufacturing,2016,29(4):283-291.[14]Jian Wan, Seán McLoone.Gaussian proc-ess regression for virtual metrology enabled run-to-run control in semiconductor manu-facturing[J].IEEE Transactions on Semi-conductor Manufacturing,2018,31(1):12-21.[15]Toshiya Hirai, Manabu Kano. Adaptive virtual metrology design for semiconductor dry etching process through locally weighted partial least squares[J].IEEE Transactions on Semiconductor Manufactur-ing,2015,28(2):137-144.

相似文献/References:

[1]卢 菲,高振斌.基于小波变换的QRS波群检测[J].电子设计工程,2011,(23):46.
 LU Fei,GAO Zhen-bin.QRS wave detection based on wavelet transform[J].SAMSON,2011,(01):46.
[2]路 翀,刘晓东,刘万泉.一种2DDCT与压缩感知结合的人脸识别[J].电子设计工程,2011,(21):186.
 LU Chong,LIU Xiao-dong,LIU Wan-quan.A face recognition algorithm based on combination 2DDCT and compressed sensing[J].SAMSON,2011,(01):186.
[3]石小云.基于SOM神经网络的静止卫星云图分类[J].电子设计工程,2011,(16):55.
 SHI Xiaoyun.Stationary satellite image classification based on SOM neural network[J].SAMSON,2011,(01):55.
[4]童 啸.分块NSA在人脸识别上的应用[J].电子设计工程,2011,(15):156.
 TONG Xiao.Modular NSA?蒺s application to face recognition[J].SAMSON,2011,(01):156.
[5]陈 胜.一种M2DPCA和NFA相结合的人脸识别方法[J].电子设计工程,2011,(13):163.
 CHEN Sheng.A face recognition algorithm based on combination of modular 2DPCA and NFA[J].SAMSON,2011,(01):163.
[6]刘 扬,王 彬,韩 雷.基于卫星云图的云分类研究[J].电子设计工程,2011,(10):189.
 LIU Yang,WANG Bin,HAN Lei.A review study of cloud classification using satellite imagery[J].SAMSON,2011,(01):189.
[7]孙艳丽,李建海,周 伟,等.SIFT在高分辨率SAR图像自动配准中的性能分析[J].电子设计工程,2011,(07):180.
 SUN Yan-li,LI Jian-hai,ZHOU Wei,et al.Performance analysis of SIFT to registration of high-resolution accuracy SAR imagery[J].SAMSON,2011,(01):180.
[8]张 皓,刘纪元,焦学峰,等.自适应峰值匹配法在风速测量中的应用[J].电子设计工程,2011,(02):69.
 ZHANG Hao,LIU Ji-yuan,JIAO Xue-feng,et al.Adaptive peak matching method in the ultrasonic wind speed measurement[J].SAMSON,2011,(01):69.
[9]阮 雷,叶玉堂,王鼎元,等.基于RBF的纸币序列号识别方法[J].电子设计工程,2010,(09):51.
 RUAN Lei,YE Yu-tang,WANG Ding-yuan,et al.Series number recognition of paper currency based on RBF[J].SAMSON,2010,(01):51.
[10]贾花萍.思维状态的EEG信号分类方法研究[J].电子设计工程,2010,(06):118.
 JIA Hua-ping.Classification for EEG signals of different mental tasks[J].SAMSON,2010,(01):118.

备注/Memo

备注/Memo:
收稿日期:2019-05-20 稿件编号:201905090作者简介:周江兵(1995—),男,湖北黄梅人,硕士。研究方向:大数据分析与数据挖掘。
更新日期/Last Update: 2019-12-30